1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
name: CI
# CI validates integration and performance on every push to master/main
# and on every pull request. Unit tests run only on PRs and on non-release
# direct pushes to master (release commits delegate to release.yml).
#
# Build architecture (Option A — parallel):
#
# build-binaries ──┬→ test-e2e (needs: build-binaries + test-unit)
# └→ test-performance (needs: test-e2e)
# test-unit ───────┘
#
# `build-binaries` compiles release binaries ONCE and uploads them as a
# workflow artifact. `test-e2e` and `test-performance` download the
# pre-built binaries and pass KNOT_SKIP_BUILD=1 to the test scripts so
# they do NOT rebuild. `test-unit` runs its own debug/test build in
# parallel (different target profile, not worth sharing).
on:
push:
branches:
pull_request:
branches:
env:
CARGO_TERM_COLOR: always
RUST_BACKTRACE: 1
jobs:
build-binaries:
name: Build Release Binaries
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install Rust
uses: dtolnay/rust-toolchain@stable
- name: Cache cargo registry
uses: actions/cache@v4
with:
path: ~/.cargo/registry
key: ${{ runner.os }}-cargo-registry-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo index
uses: actions/cache@v4
with:
path: ~/.cargo/git
key: ${{ runner.os }}-cargo-index-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo build
uses: actions/cache@v4
with:
path: target
key: ${{ runner.os }}-cargo-build-target-${{ hashFiles('**/Cargo.lock') }}
- name: Build release binaries
run: cargo build --release --all-features
- name: Upload release binaries
uses: actions/upload-artifact@v4
with:
name: knot-binaries
path: |
target/release/knot-indexer
target/release/knot
target/release/knot-mcp
retention-days: 1
if-no-files-found: error
test-unit:
name: Unit Tests & Linting
if: github.event_name == 'pull_request' || (github.event_name == 'push' && !startsWith(github.event.head_commit.message, 'release:'))
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install Rust
uses: dtolnay/rust-toolchain@stable
- name: Cache cargo registry
uses: actions/cache@v4
with:
path: ~/.cargo/registry
key: ${{ runner.os }}-cargo-registry-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo index
uses: actions/cache@v4
with:
path: ~/.cargo/git
key: ${{ runner.os }}-cargo-index-${{ hashFiles('**/Cargo.lock') }}
- name: Format Check
run: cargo fmt -- --check
- name: Clippy
run: cargo clippy --all-targets -- -D warnings
- name: Run Unit Tests
run: cargo test --lib --all-features
test-e2e:
name: E2E Integration Tests
runs-on: ubuntu-latest
needs:
- build-binaries
- test-unit
# Run if build succeeded AND (test-unit succeeded OR was skipped).
# test-unit is skipped on release: commits, but e2e should still run.
if: always() && needs.build-binaries.result == 'success' && (needs.test-unit.result == 'success' || needs.test-unit.result == 'skipped')
steps:
- uses: actions/checkout@v4
- name: Install Rust (for any auxiliary cargo invocations)
uses: dtolnay/rust-toolchain@stable
- name: Download pre-built binaries
uses: actions/download-artifact@v4
with:
name: knot-binaries
path: target/release
- name: Make binaries executable
run: chmod +x target/release/knot-indexer target/release/knot target/release/knot-mcp
- name: Cache cargo registry
uses: actions/cache@v4
with:
path: ~/.cargo/registry
key: ${{ runner.os }}-cargo-registry-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo index
uses: actions/cache@v4
with:
path: ~/.cargo/git
key: ${{ runner.os }}-cargo-index-${{ hashFiles('**/Cargo.lock') }}
- name: Cache fastembed models
uses: actions/cache@v4
with:
path: ~/.cache/knot/fastembed_cache
key: ${{ runner.os }}-fastembed-v2
- name: Pre-download fastembed model (with retries for 429 rate limit)
run: |
# Create cache dir
mkdir -p ~/.cache/knot/fastembed_cache/fast-bge-small-en-v1.5
mkdir -p ~/.cache/knot/fastembed_cache/all-minilm-l6-v2
# Only download if missing (saves time if cache restored successfully)
if [ ! -f ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/model.onnx ]; then
echo "Model not in cache. Downloading with retries..."
# Try up to 5 times with exponential backoff
for i in {1..5}; do
# We use AllMiniLML6V2 as default in tests
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/model.onnx" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/model.onnx &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/config.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/config.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/tokenizer.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/tokenizer.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/tokenizer_config.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/tokenizer_config.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/special_tokens_map.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/special_tokens_map.json && break || {
echo "Download failed (attempt $i). Waiting..."
sleep $((2**i))
}
done
# Write a marker so fastembed knows it's the right model
echo "all-MiniLM-L6-v2" > ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/.fastembed_model_name 2>/dev/null || true
else
echo "Model found in cache."
fi
- name: Install dependencies (netcat, time)
run: sudo apt-get update && sudo apt-get install -y netcat-openbsd time
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Run all E2E tests (Fast)
env:
KNOT_SKIP_BUILD: "1"
run: ./tests/run_all_e2e_fast.sh
timeout-minutes: 60
- name: Upload E2E logs on failure
if: failure()
uses: actions/upload-artifact@v4
with:
name: e2e-logs
path: |
tests/.e2e_data/
tests/.e2e_*/
retention-days: 7
test-performance:
name: Performance Benchmarks
runs-on: ubuntu-latest
needs: test-e2e
if: success()
steps:
- uses: actions/checkout@v4
- name: Install Rust (for any auxiliary cargo invocations)
uses: dtolnay/rust-toolchain@stable
- name: Download pre-built binaries
uses: actions/download-artifact@v4
with:
name: knot-binaries
path: target/release
- name: Make binaries executable
run: chmod +x target/release/knot-indexer target/release/knot target/release/knot-mcp
- name: Cache cargo registry
uses: actions/cache@v4
with:
path: ~/.cargo/registry
key: ${{ runner.os }}-cargo-registry-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo index
uses: actions/cache@v4
with:
path: ~/.cargo/git
key: ${{ runner.os }}-cargo-index-${{ hashFiles('**/Cargo.lock') }}
- name: Cache fastembed models
uses: actions/cache@v4
with:
path: ~/.cache/knot/fastembed_cache
key: ${{ runner.os }}-fastembed-v2
- name: Pre-download fastembed model (with retries for 429 rate limit)
run: |
# Create cache dir
mkdir -p ~/.cache/knot/fastembed_cache/fast-bge-small-en-v1.5
mkdir -p ~/.cache/knot/fastembed_cache/all-minilm-l6-v2
# Only download if missing (saves time if cache restored successfully)
if [ ! -f ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/model.onnx ]; then
echo "Model not in cache. Downloading with retries..."
# Try up to 5 times with exponential backoff
for i in {1..5}; do
# We use AllMiniLML6V2 as default in tests
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/model.onnx" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/model.onnx &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/config.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/config.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/tokenizer.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/tokenizer.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/tokenizer_config.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/tokenizer_config.json &&
curl -fL "https://huggingface.co/Qdrant/all-MiniLM-L6-v2-onnx/resolve/main/special_tokens_map.json" -o ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/special_tokens_map.json && break || {
echo "Download failed (attempt $i). Waiting..."
sleep $((2**i))
}
done
# Write a marker so fastembed knows it's the right model
echo "all-MiniLM-L6-v2" > ~/.cache/knot/fastembed_cache/all-minilm-l6-v2/.fastembed_model_name 2>/dev/null || true
else
echo "Model found in cache."
fi
- name: Install dependencies (netcat, time)
run: sudo apt-get update && sudo apt-get install -y netcat-openbsd time
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Run Performance Benchmarks
env:
KNOT_SKIP_BUILD: "1"
run: |
./tests/benchmark_e2e.sh --focus java_e2e --output-dir /tmp/perf_results
scripts/compare_perf_metrics.sh /tmp/perf_results .perf_metrics/baseline.json
timeout-minutes: 10
- name: Upload performance metrics
if: always()
uses: actions/upload-artifact@v4
with:
name: perf-metrics
path: /tmp/perf_results/
retention-days: 30
- name: Update Baseline (on main/master merge)
if: github.ref == 'refs/heads/main' || github.ref == 'refs/heads/master'
run: |
LATEST_AGG=$(find /tmp/perf_results -name aggregated.json -type f | sort | tail -1)
if [ -n "$LATEST_AGG" ]; then
cp "$LATEST_AGG" .perf_metrics/baseline.json
echo "Baseline updated for $(git rev-parse --short HEAD)"
else
echo "WARNING: No aggregated.json found in /tmp/perf_results"
fi